Proceedings of the 26th Nordic Seminar on Computational Mechanics

A. Logg, K.A. Mardal (Eds.)

c
Oslo, 2013

Modelling ship grounding with finite elements

Ole J. Hareide1 , Lars Brubak1,2 , and Torstein Pettersen1

(1) Section
for Ship Structures and Concepts, Maritime Advisory, Det Norske Veritas, Norway.
E-mail: ole.jakob.hareide@dnv.com
(2) Mechanics Division, Department of Mathematics, University of Oslo, Norway

Summary. Numerical simulations using non-linear finite element analysis have been perform to study
ship grounding. The results are compared with experimental tests which was performed in USA in 1995
and good agreement is achieved. This illustrates that nonlinear FE analysis can be used to estimate
damage extent from a grounding scenario.

Key words: Ship grounding, FEA, non-linear analysis

Introduction
Every year ships run aground around the world. The potential for damage is huge, both in
economic and ecological terms. With the ability to simulate ship groundings with finite element
software and assess the strength of a ship against such incidents there is a huge potential to
be gained. Similar studies for collision scenarios by using nonlinear finite element analysis have
been performed by Notaro et. al [2, 1]
A detailed picture of a ships performance in a grounding situation can be used in the design
of the ship. The results do, however, depend on correct material parameters and the use of a
correct coefficient of friction in addition to a very detailed description of the geometry. There
is not much data to be found from real events so to verify that the model is indeed giving
reasonable results, a series of analyses has been run against results obtained from scale tests
performed in the USA in 1995 [4].

Background
A series of model scale test was performed at the Naval Surface Warfare Center in the US in
1995. Four different steel configurations of double bottom structures were built in 1:5 scale and
used to evaluate the performance when impacted into a concrete cone used to simulate a rock
on the seabed. The models were mounted to a rail cart and released down a slight incline in
order to build up velocity and hit the rock at a speed of 14 knots. In a double hull configuration
there is generally not a danger for a cargo spill if the inner hull is not breached, so the double
bottom models were mounted at a pitch angle of approx. 3.2 in the test rig to ensure a gradual
vertical increase in the damage depth and eventually a rupture of the inner hull. The output
from the experiments, which have been used to compare the finite element results against, were
force-displacement and energy-displacement curves.

FE model and analysis procedure

A detailed model of one of the double bottom scale models was created in Abaqus/Explicit v6.11
[3]. Of the four models tested the one chosen for modelling in Abaqus was the configuration
 (a) (b)
Figure 1. (a) A barge impacting into a rock and (b) the detail level and mesh for the barge where
stiffeners, girders and bulkheads are included.

most resembling the double bottom of a traditional tanker as shown in Fig. 1. The model
consisted of two cargo holds divided by a stiffened bulkhead. On either end of the hold there
were also stiffened bulkheads. The double bottom consisted of a longitudinal girder in the centre
and seven transverse frames under each hold. The rock was modelled as a rigid body with a
cone shape.

First-order reduced-integration shell elements were used with a fine mesh size of around 10x10mm
in a fine area where the rock would impact and 50mm in the rest of the model. The reason for
such a fine mesh, which is in the order of three times the thickness, is to accurately capture the
failure mechanisms such as rupture, folding and crushing of plates and stiffeners.
The material used in the model was an ASTM A569 isotropic steel with a Youngs modulus of
E = 206000MPa, Poissons ratio of = 0, 3 and density of = 7.9 103 kg/m3 . Plasticity was
included with a yield stress of y = 283MPa. In order to model the rupture of the steel as the
rock passed through the model a tensile failure criterion was added to the material definition
which after tensile strain of 0,278 will simulate necking behaviour in the material and linearly
degrade an elements stiffness down to zero over a damage displacement, defined here as 3,2mm
additional displacement after necking occurs. When the elements stiffness has reached zero it
is visually removed from the analysis.
The coefficient of friction is difficult to measure from a grounding experiment due to other en-
ergy dissipating mechanisms such as tearing and folding of the structure. Typical values of steel
against rock ranging from 0.4 to 0.7 have been used here, and the choice of this coefficient can
have a large impact on the dissipation of energy.
The analyses were run with Abaqus/Explicit with an initial velocity of 14 knots applied to the
model. Along with the mass of the structure an additional mass was added to the model to
account for the weight of the testing rig used in the scale model tests. Using an initial kinetic
energy (as opposed to the constant velocity) approach means that the kinetic energy will be
dissipated by the impact and that the model will come to a halt when the dissipated energy
equals the initial kinetic energy. The upper two longitudinal edges in the structure were fixed
in order to keep them rigid during the analysis.

Results and discussion

The results from the analyses are shown here in Fig. 2. Both energy and force curves are plotted
against the displacement of a reference point placed in the rear of the model. The kinetic energy
 (a) (b)
Figure 2. (a) Reaction force and (b) absorbed energy versus the grounding distance.

(a) (b)
Figure 3. (a) Damage extent of the barge and (b) close-up view of the damages.

will drop as the energy dissipation increases. From the force curve, several smaller spikes and
the two large spikes are observed, corresponding to the structure resisting deformations as the
rock passes through the several smaller transverse frames and the two cargo hold bulkheads at
the center and the aft. There is also an overall increasing trend in the registered force which is
due to the pitch angle. The absorbed energy can be found by integrating the reaction force over
the grounding distance and thus it can be seen to also have a slightly higher absorption rate
(energy absorption pr. meter) around the two bulkheads than in the cargo hold itself.
The total dissipated energy can be broken down into several components such as friction, plastic
deformation, elastic strain and energy gone into tearing elements apart. In the analyses that
were run it was seen that most of the energy is dissipated by plastic deformation and friction
while the other aforementioned energies that make up the rest are relatively small. The ratio
between the energy going into friction and plastic deformation varies depending on the coefficient
of friction, as was mentioned above.
The damages are shown in Fig. 3 and it can be seen that the deformations are very large. This
illustrate that nonlinear finite element computations can be used to estimate the damage extent
with reasonable accuracy. From what can be read from the pictures of the experimental test
[4], the damages are very similar to what was found in the present finite element analysis. In
Fig. 4, close-up views of the deformations in the longitudinal girder and in the area close to the
transverse bulkhead are shown.
 (a) (b)
Figure 4. Close-up view of (a) the deformation in the longitudinal girder which is a typical skewed folding
pattern from shear deformations and (b) the deformations at the transverse bulkhead.

Conclusion
Finite element analysis for a grounding scenario has been performed and computed results
are compared and verified with experimental test data from a steel 1:5 scale model. Results
show that the simulation can accurately predict the damage extent and yield the same energy
dissipation rate.

Acknowledgments
The authors would like to thank dr. Eivind Steen, Tom K. stvold, Gabriele Notaro and Alessio
Prestileo at Det Norske Veritas (DNV), for their interest, suggestions and valuable discussions
throughout the study.

References
[1] G.Notaro, K. Brinchmann, E. Steen and N. Oma. Evaluation of the Fendering Capabilities
of the SPS for an Offshore Application, ICCGS 2013 Conference, Trondheim, Norway, 2013.

[2] G.Notaro, T.K. stvold, E. Steen and N. Oma. Collision Damages and Residual Hull Girder
Strength of a Ship Shaped FPSO, PRADS 2013 Conference, Changwon City, Korea, 2013

[3] ABAQUS/Standard, User Manual Version 6.11 volume Hibbitt, H. D. and Karlsson, B. I.
and Sorensen, I. Hibbit, Karlsson and Sorenson Inc., Pawtucket; 2012.

[4] James L. Rodd. Observations on Conventional and Advanced Double Hull Grounding Experi-
ments Proceedings of the International Conference on Design and Methodologies for Collision
and Grounding Protection of Ships. San Francisco, California, 1996